Measurements, characterization, and system design for digital storage
In this dissertation we examine the properties of new materials for high density tape recording and develop new signal processing techniques for digital magnetic recording. The first three chapters discuss tape recording research, while the latter chapters discuss signal processing techniques for digital recording.
Chapter 2 details the construction of a device for testing digital tape recording media. This device was designed as part of the National Storage Industry Consortium (NSIC) program for high density tape recording and provides a simple interface for measuring the recording properties of tape media. Chapter 3 discusses a variety of measurements taken for the (NSIC) tape program. These measurements characterize some of the new tape media with thin magnetic layers that are being produced today. The experimental work was performed both at UCSD and at the Eastman Kodak Company in San Diego. Chapter 4 describes a study to determine the performance of the new, thin tape media with regard to high density tape recording. It also compares the performance of the new media to traditional tape media, which has a much thicker magnetic recording layer.
Chapter 5 analyzes the ternary partial response channel with channel polynomial $h(D) = 1-D$. This chapter examines the error events and probability of error for a ternary system. In Chapter 6, a simple implementation of the Viterbi algorithm for the ternary $1-D$ channel is developed. It is shown that this implementation can be implemented in analog fashion and is simpler, in terms of the number of operations required, than the standard Viterbi algorithm. Chapter 7 describes a new implementation of the extended class 4 partial response (EPR4) channel with channel polynomial $h(D) = (1-D)(1+D)\sp2$. This new detector uses two ternary $1-D$ Viterbi detectors as its main detection unit. Finally, in Chapter 8, a concatenated detection algorithm for simplifying the detection of distance-enhancing trellis codes is introduced. An example of the algorithm is given for a trellis code on the EPR4 channel.